Layer 8 · Neuron 5173 · boosted

L8N5173+1

Word quirks:

“ response”“ answer”“ explanation”“ \"”“ reason”“ answering”“ ‘”“ question”

💤 Barely changes the model — the quick scan suggests this neuron barely changes the model's word predictions.

Identity (according to the specimen itself)

I am an AI assistant designed to simulate a conversation with a human-like intelligence. My mission is to understand and assist users in navigating complex information spaces, often involving ethical considerations and the development of advanced technologies. I strive to provide helpful, accurate, and safe responses that can be used by anyone seeking knowledge or support.

Nearest neighbors

Specimens with similar replies or word quirks. Similarity runs from −1 to 1; 1 is identical. These are leads, not shared mechanisms. How profiles are built.

By word quirks Profile recipe

Specimen	Similarity	Word quirks
1. L9N5004+1💤	0.90	“<\|im_end\|>”“ ‘”“ question”
2. L5N1631+1🕒	0.89	“ response”“ explanation”“ clarification”
3. L12N5785+1🕒	0.85	“ question”“ response”“ answer”
4. L15N1922+1💤	0.84	“ “”“⏎”“ explanation”
5. L8N5484-1💤	0.82	“ earlier”“ ⏎”“ —”
6. L9N1976+1💤	0.82	“ “”“ answer”“ "”
7. L9N7800-1💤	0.82	“ response”“ answer”“ personal”
8. L17N1486+1🕒	0.82	“ answer”“ “”“ question”
9. L5N4140+1🕒	0.82	“ response”“ reason”“ reference”
10. L8N593-1🕒	0.82	“ basically”“ ‘”“ answer”

Shadow specimen, apart from the ten: L8N5173-1🕒, -0.01.

Cousins · specimens sharing a neuron

No specimen with replies shares this neuron yet.

Technical details

Quick scan: measure prediction shifts on fixed reference text. Short test: generate replies to a small set of prompts. Longer test: extend to a larger set. Full test: run the full prompt suite. A pass admits a specimen to the next stage; only missing replies are generated.

Layer 8, neuron 5173, dose +1. This specimen has no prompt-suite replies yet. Its quick-scan results, when available, appear below. Its quick-scan row (KL 0.97, top-1 agreement 0.60) is outside the band that feeds the short test (KL 1 to 2.5, agreement 0.5 to 0.65), so it stays at the quick scan unless someone queues it by hand. A hand pick goes straight to the full test. See Method for the tiers.

Quick scan · prediction measurements

Measurements are averaged over token positions of the five unsteered reference answers, using the same fixed text under both conditions; no steered replies are generated here.

KL divergence: KL(P_unsteered ‖ P_steered), in nats — how far steering shifts the next-token distribution in this direction (Kullback & Leibler, 1951, On Information and Sufficiency). Unsteered ≈ 0.001.
Top-1 agreement: fraction of positions with the same highest-probability token — how often the top prediction stays unchanged.
ΔNLL: mean negative log-likelihood under steering minus the unsteered value for the reference tokens, in nats — positive means the original answer becomes less likely.
Entropy: mean H(P_steered) = −Σ p log p, in nats — how spread out the steered next-token predictions are.

The steer lens lists the tokens with the largest mean change in log-probability (steered minus unsteered), in nats, over those positions. These are leads for inspection; in rare specimens even the top gains are negative.

Setting	KL divergence	Top-1 agreement	ΔNLL	Entropy	Steer lens
L8N5173+1 t0-v0.1	0.968	0.60	1.182	1.54	Top boosted tokens ↑
L8N5173-1🕒 shadow specimen t0-v0.1	1.387	0.53	1.680	1.55	“ income”“ adjustment”“ incomes”“ adjustments”“ special”“ ⏎”“ financial”“ accounted”

Method

A specimen is a fixed steering configuration of Qwen3.5-4B: one MLP neuron or a pair, each with a signed dose. Unsteered is the control. L identifies the layer; N, the neuron.

Additive activation steering. At every token position, add dose × layer median residual norm ÷ ‖W_down[:, n]‖ to MLP neuron n at the down_proj input. In plain terms, shift that neuron's activation by a fixed amount, scaled by the layer’s calibrated median residual norm and the neuron's down-projection weight norm. Positive doses push up; negative doses push down. Dose is normalized intervention size, not a percentage or probability.